Evidence suggesting an association between n-3 polyunsaturated fatty acids (PUFA) and stroke risk has been inconsistent, possibly because previous studies have not differentiated between different stroke types. The present study investigated the hypothesis that tissue levels of n-3 PUFA are positively associated with hemorrhagic stroke and negatively associated with ischemic stroke. We recruited 120 subjects for this case-control study, with 40 cases each of hemorrhagic stroke, ischemic stroke, and unaffected controls. Patients with a family history of hemorrhagic stroke had a significantly increased risk for hemorrhagic stroke. Omega-3 Index (20:5n3 + 22:6n3 in erythrocytes) and 22:6n3 were negatively (P < .01) associated with the risk of both hemorrhagic and ischemic stroke in multivariate analyses. Saturated fatty acids 16:0 and 18:0 were positively associated, whereas 18:2n6 and 18:3n6 were negatively (P < .05) associated with risk of ischemic stroke. Monounsaturated fatty acid, 18:1n9, increased (P = .03) the odds of hemorrhagic stroke. Omega-3 Index and docosahexaenoic acid were significantly lower in patients with both subtypes of hemorrhagic stroke, subarachnoid and intracerebral hemorrhage, but only in one subtype of ischemic stroke, small-artery occlusion. Saturated fatty acids 16:0 and 18:0 were significantly higher, but 20:4n6 was significantly lower, in patients with small-artery occlusion. Linoleic acid was significantly lower in patients with small-artery occlusion and large-artery atherosclerosis, whereas 18:1n9 was higher in both subgroups of hemorrhagic stroke. In conclusion, the results of our case-control study suggest that erythrocyte n-3 PUFA may protect against hemorrhagic stroke and ischemic stroke, particularly in the case of small-artery occlusion.